About Objective Partners
Objective Partners is a young and dynamic marketing analytics company. We offer a combination of software and consulting services to help advertisers optimize their media investments. As we are growing and expanding quickly, we are looking for graduation trainees to join our awesome team!
What do we offer?
• A graduation traineeship, combining 3 days of thesis and 2 days of work
• A clear thesis subject with practical relevance
• All datasets are ready before you start
• Guidance from O/P’s experienced data scientists
• Getting the unique opportunity working for the top advertisers in Europe
• The chance to be a part of a fast-growing and ambitious company
• Awesome and regular team outings, sports, dinners and trainings
Interested? Send your application to firstname.lastname@example.org
Want to know more about what we do or about working at Objective Partners? Take a look at our website and job openings here.
1. Media Attribution Using Machine Learning
Within O/P we have implemented multiple attribution models to calculate the added value of online media investments. The model we use at our customers is based on Game Theory, to be more specific on the Shapley value which is discussed in the paper by Dalessandro . We are constantly looking for ways to improve our attribution model and exploring different techniques like Neural networks. In the paper by Ren  a Recurrent Neural Network for the multi-touch attribution problem is proposed. In your thesis you will be implementing this or a similar technique and compare the results to our current model.
Want to know more about attribution models? Read one of our blogs about multi-touch attibution.
 Causally motivated attribution for online advertising (B Dalessandro, et al)  Learning Multi-touch Conversion Attribution with Dual-attention Mechanisms for Online Advertising (Ka Ren, et al)
2. Bayesian Methods for Media Mix Modeling and Budget Allocation
Within O/P we have implemented Media Mix Modelling (MMM) to calculate the effects of TV and Radio (and other media channels) on bottom line sales. One of the challenges with a media mix model are the carryover and shape eﬀects of advertising. Next to this, media managers often have a strong hypothesis (prior knowledge) on the effect of a certain channel on bottom line performance (for example based on the number of people reached by that channel). In your thesis you will investigate the Bayesian Methods for Media Mix Modeling proposed by  and possibly extend this approach to be working in practice. This approach captures both the challenge carryover and shape eﬀects and has room for the use of prior knowledge.
 Bayesian Methods for Media Mix Modeling with Carryover and Shape Eﬀects (Y Jin, et al)